Consistent inspection process

An automated ‘green button’ manufacturing process could reduce reliance on offline part inspections. Dave Wilson reports.


Most manufacturers currently use offline part measurement systems, such as co-ordinate measuring machines (CMMs), to verify the dimensions of high-value products.

But despite their high accuracy, making such measurements with CMMs can be a time-consuming process that must be performed in specialised environments removed from the production floor, resulting in increased production lead times and manufacturing costs.

In tandem with this, many firms measure parts manually at intervals during metal cutting for process-control purposes. As with any manual process, this can come with its own inherent variation and risk of human error.

A growing school of thought sees the new Holy Grail of manufacturing as reducing reliance on offline part inspection by using data collected prior to and during manufacturing to ensure that any process produces the consistent results that are required. The challenge is to do this in an automated, repeatable fashion to deliver a ‘green button’ process — one where the operator presses the cycle start button and walks away.

According to Marc Saunders, general manager of group marketing services and UK Sales at precision engineering group Renishaw, to reach that manufacturing nirvana, careful attention must be paid to each stage of the manufacturing process.

Prior to machining any part, a machine tool must be in a suitable condition to make the part with sufficient accuracy and consistency. For that reason, calibration and condition monitoring systems must be deployed at regular intervals.

On an annual basis, laser inter- ferometer systems should be used to map out the residual inaccuracies in a machine’s structure. More regularly, devices such as telescoping ball bars should be deployed, typically on a weekly or monthly basis, to monitor the condition of the machine to ensure that it has not degraded. A ball bar can identify axis squareness, reversal spikes, backlash, servo mismatch and various other error sources to ensure that a machine is in a consistent, healthy state.

‘Such procedures provide manu- facturers with the confidence that the manufacturing process will be repeatable and that they will be able to make the part in tolerance’, said Saunders.

The next step is to set up the part and cutting tools correctly on the machine tool prior to machining. While this was once a predominantly manual process, today, it is common to use measurement and tool-setting probes to automate the procedure.

Measurement probes, which are mounted in the spindle of a machining centre or the turret of a lathe, can be used for a number of set-up functions, including determining the datum position on the work piece, as well as its angular alignment. Meanwhile, tool setting probes can determine the lengths and diameters of the cutting tools that will be used to make the part.

This is not only beneficial from a productivity point of view, it is also a more repeatable procedure than ‘cut and measure’ techniques. It also avoids errors resulting from manually keying in offsets, which are one of the major causes of crashes and tool breakage in many machine shops.

According to Saunders, while calibration and set-up are important, the need to acquire data during machining and make use of it to optimise the manufacturing processes in situ is also now recognised as a vitally important area, especially where the machining of a part is complex and time consuming.

In such processes, many of which are found in the aerospace industry, tool wear is high due to the properties of the materials that are being machined and many tools may be consumed during the manufacture of a single part.

‘Although initial offsets are established when the tool is set up, either by a contact or non-contact laser system, the offsets will need to be changed during the life of the tool, as wear causes the part to drift out of tolerance,’ said Saunders.

Where tolerances are tight and exotic materials are used, as is the case in the aerospace industry, tools will often wear to a greater physical degree than the tolerance of the feature that manufacturers are trying to make.

During the machining process, it is the part that must be monitored and any small errors in the measured size of the part used to tweak the tool offset to keep the feature as close to nominal as possible as the tool becomes progressively worn over the course of its life.

At present, some manufacturers use probes that are mounted in the spindle of the milling machine to perform such operations. The probe is usually on a shank that has some form of signal transmission system with which it can communicate with the machine tool’s controller. The probe gathers data points across the feature of the part, using software on the CNC to compare with the nominal size to automatically adjust the tool set appropriately.

An alternative up-and-coming candidate for such in-situ measurement and control is fringe projection. Fringe projection systems provide a long working distance and can hence withstand harsh machining environments. They are flexible to set up and are inherently robust since they have no moving parts. They do, however, require an optical co-operative surface and may struggle to access internal features.

Typically, in the manufacture of high-value parts where the material cost may be many thousands of pounds, manufacturers do not make a finished component in a single pass. Often, an initial roughing process is followed by several finishing processes that may be performed using a different tool.

In between each process, measurements are performed to ensure that the final cut will meet the tolerances demanded by the manufacturer.

‘Many aerospace manufacturers have traditionally taken these intermediate measurements manually using large micrometers. But due to the increasingly tight tolerances of their parts, they are looking to automate this procedure too,’ said Saunders.

To do so, they are also aiming to make automated in-situ measurements at both the roughing and finishing stages and using that data in a closed-loop process-control system to control the manufacturing process. Saunders added that, while such techniques are very important, the frequency with which they need to be deployed must be considered on a case-by-case basis due to the specific amount of machining that needs to be made and the rate of wear on the tools in any given process.

One last important process parameter that needs to be considered is the effect of changes in temperature, which can be especially important when machining a part with longer cycle times. For large aerospace parts, temperature variations can significantly affect the tolerance of the part that is being machined.

‘Fortunately, checking a machine periodically with a probe is an uncomplicated procedure that allows users to recalibrate the machine and ensure that they are taking account of the thermal condition of the machine both before and during the manufacturing process,’ said Saunders.

But such thermal effects must also be considered while a part is being manufactured. The thermal errors of a spindle, for example, have a significant effect on machining accuracy and directly influence the shape of the finished part. Fortunately, capacitance sensor systems can here be used to measure this effect and compensate for the errors that may be incurred.

With such a plethora of measuring systems in or soon to be in place, could it be that the traditional CMM will no longer have a place in manufacturing?

Saunders doesn’t think so. He believes that while in-situ process monitoring systems may ultimately reduce the need for offline part measurement systems to control the accuracy of the manufacturing process, they will still be widely used in the verification process to validate designs.

With the recent introduction of five-axis sensor systems that enable surface-finish measurement, as well as much faster inspection of component geometry, the CMM is set to become more productive and capable than ever before — a multi-functional device that can assess both part geometry and integrity, consolidating disparate testing processes on a single CNC platform.


Dave Wilson